Custom Architecture for Effective Semantic App Search: A Systematic Approach

Semantic App Search revolutionizes the effectiveness and precision of app search queries by leveraging semantic technologies and natural language processing techniques. Unlike traditional approaches that rely on exact matches between the words of the user query with the app name and description, this solution incorporates semantic comprehension of the text, overcoming the limitations of conventional methods. To capture the underlying semantic linkages and contextual understanding’ a custom architecture is employed that extracts useful information from app descriptions and user queries. Multiple sentence transformers along with keyword extraction algorithms have been utilized to improve the app-to-query mapping. An ablation analysis of the proposed architecture has been done using the NDCG parameter, with the architecture getting a score of 0.9152 out of 1. Through our implementation of semantic app search, users can discover and find apps with higher satisfaction and efficiency. By harnessing semantic technologies and natural language understanding, we aim to bridge the gap between user intent and app descriptions, facilitating more accurate and relevant recommendations.